by Thomas B. Hoffer, Karen Grigorian,
and Eric Hedberg, NORC
According to data from the 2006 National Science Foundation (NSF) Survey of Doctorate Recipients (SDR), 45% of the most recent science, engineering, and health (SEH) doctorate recipients had completed or were participating in postdoctoral appointments, up from 41% in 1995. The median length of postdoctoral appointment over that time period remained at approximately 2 years, continuing a 30-year trend.
Long a standard transition stage for doctorate recipients
in the life sciences, the postdoctoral appointment (postdoc) has become increasingly common in all SEH fields. The SDR defines a postdoc as a temporary position
awarded in academe, industry, a nonprofit organization,
or government primarily for gaining additional education and training in research. This InfoBrief examines
postdoc participation patterns by doctoral fields of study, focusing on patterns of stability and change between doctoral cohorts.
Trends in the Use and Duration of Postdocs
New doctorate recipients are increasingly likely to take postdocs, and that is evident in the 2006 SDR data: among all SEH doctorate recipients, 38% had held a postdoc at some point in their careers (table 1). More recent cohorts were more likely than earlier ones to have held a postdoc: 45% of those earning the doctorate
within the last 5 years compared with 31% of those who earned the doctorate more than 25 years ago.
Table 1 Source Data: Excel file
The likelihood of taking a postdoc varies by field of doctorate. Individuals who earned doctorates in the life sciences (57%) or physical sciences (50%) were most likely to report having had a postdoc. Doctorate recipients in engineering (21%), computer/mathematical sciences (21%), and the social sciences (including psychology) (23%) were much less likely to have held a postdoc.
The increases in postdoc participation rates from earlier to more recent doctoral cohorts were greatest for doctorate
recipients in engineering (164%) and computer/mathematical sciences (82%). Increases in these fields were particularly pronounced between the most recent cohort (5.0 years or less) and all earlier cohorts. Life sciences doctorate recipients had the largest absolute increase between the earliest cohort (49%) and those who earned the doctorate 20.1 to 25.0 years prior (57%) but little growth in postdoc participation among more recent cohorts. A similar pattern was evident for physical sciences doctorate recipients, except that the largest jump occurred between the 20.1–25.0 and 15.1–20.0 years-since-doctorate cohorts.
Among those who have held postdocs, the time spent in the appointment can vary substantially. Longer overall time in postdoctoral status can reflect lengthier single appointments or multiple, usually successive, postdocs. Figure 1 shows that the median duration of the most recently completed postdoc generally increased from the early 1960s until the early 1970s but remained fairly stable at about 2 years for the next 30 years. The length of postdocs varies by field of study. Postdoc holders who earned their doctorates in life sciences, physical sciences, and computer/mathematical sciences had longer postdocs than did those in engineering and social sciences (including psychology) (table 2).
Figure 1 Source Data: Excel file
Table 2 Source Data: Excel file
Employment Sector of Postdocs
Although postdocs are often equated with academic or university research settings, about one-fourth of postdoc holders are employed by the government, private nonprofit organizations, and industry. For the SEH population as a whole, 75% of those who held postdocs were employed by an educational institution, 11% worked in for-profit or nonprofit companies or organizations, and 12% worked in government (table 3). This distribution was generally similar for postdoc holders from each of the broad fields of doctoral study. The main exception was social sciences; doctorate recipients in these fields accept a higher proportion of postdocs in for-profit/nonprofit organizations than in government. Further analysis shows that this employment-sector deviation is driven by employment patterns for recipients of psychology doctorates, 60% of whom took postdocs in academe and 24% of whom had postdocs at for-profit or other nonprofit organizations. In contrast, other social sciences doctorate recipients with postdocs were similar to postdoc holders in other fields, with 82% of the other social science doctorates working in postdocs in educational institutions and 8% in for-profit or other nonprofit organizations. Analysis of the sector locations by years since earning the doctorate show few differences among cohorts in terms of the balance between higher education, industry, nonprofit organizations, and government.
Table 3 Source Data: Excel file
Work Activities and Reasons for Taking Postdocs
Individuals who were working in a postdoc on the April 1 reference date of the 2006 SDR were asked to indicate
their primary and secondary work activities. Table 4 shows that about 88% of the current postdocs in 2006 identified research and development as a primary or secondary work activity, and 9% identified teaching. Among postdocs in 1995, 95% identified R&D as a primary or secondary activity, suggesting changes in the past decade in the work activities that constitute postdoc assignments. It does not appear that teaching as a primary or secondary work activity, however, has increased, indicating that other types of work activities are the reason for the decline in time spent on R&D.
Table 4 Source Data: Excel file
An important set of research questions concerns the reasons for taking postdocs, how those reasons may differ by discipline, and whether they are changing across doctoral cohorts. Of particular concern has been whether the growth of postdocs reflects a relative lack of tenure-track openings, such that postdocs provide a sort of safety net for individuals committed to academic careers until the preferred positions become available.
The SDR shows that the most frequently cited reason for the current or most recent postdoc was to gain additional
training in the doctoral field, identified as the main reason by 34% of all postdoc participants (table 5). This was followed by "work with a specific person or place" (19%) and "postdoc generally expected for a career in this field" (19%). Examining the main reasons for taking postdocs, a few notable differences are evident. Engineering doctorate recipients were much more likely than others to identify "other employment options not available" (23%) and were less likely than others to identify "postdoc
generally expected for a career in this field" (7%).
Table 5 Source Data: Excel file
Trends in Health and Retirement Benefits for Postdoc Participants
The availability of health and retirement benefits to postdoc participants has been an important issue to advocacy groups in recent years as the use of postdocs has grown and become a standard career step in more disciplines. The availability of both types of benefits to postdocs has increased steadily over the past several
years, with about 90% of the most recent doctoral cohort members having health benefits and about 50% having retirement benefits (figure 2). Examination of the availability of benefits by broad field of doctoral study shows that social sciences doctorate recipients (including psychology) were the least likely to have both types of coverage (table 6).
Figure 2 Source Data: Excel file
Table 6 Source Data: Excel file
Employment Sector after the Postdoc
The SDR data also allow for examination of the employment
sector where postdocs subsequently continue
their careers. Do postdocs mainly transition into academic employment, or do they find employment in industry or government instead? How have the relative shares going to each employment sector changed over time? As a preliminary approach to these questions, the current (2006) employment sector for all who reported completing at least one postdoc is shown in table 7. As a general pattern, about half of all former postdocs were employed in the education sector (almost entirely university and college appointments) (49%), but a large number were employed in the business/industry sector (41%). These percentages were largely similar across the doctoral cohorts, thus not indicating any clear trends into or out of the three broad employment sectors.
Table 7 Source Data: Excel file
Data Comments and Availability
The SDR is a biennial survey of doctorates earned in science, engineering, and health fields at U.S. institutions.
The focus of the survey is the labor force experiences
of this population and how those experiences change over the course of individual careers and across historical cohorts. A special module of questions on past postdoctorate experiences in the 1995 SDR was repeated in the 2006 SDR, and this InfoBrief focuses on the 2006 module. The 1995 and 2006 modules included questions on the total number of postdocs held and asked for detailed information on three past and present postdoc appointments, including employment sector of each postdoc, reasons for taking each postdoc,
and availability of benefits. The full set of detailed tables from the 2006 SDR will be available in the forthcoming report Characteristics of Doctoral Scientists
and Engineers in the United States: 2006, at http://www.nsf.gov/statistics/doctoratework/.
The 2006 SDR was administered to a nationally representative sample of about 40,000 SEH doctorate recipients from U.S. universities and who were residing in the United States. The sample frame for the SDR is built from the Doctorate Records File (DRF), a comprehensive
list of all research doctorate recipients and their demographic characteristics from U.S. universities from 1920 to the present. Since 1958 the DRF has been updated annually with data from the Survey of Earned Doctorates (SED), a census of new doctorate recipients sponsored by NSF and five other federal agencies. The SDR sample is augmented every 2 years with members of the new U.S. doctoral cohorts surveyed by the SED, and sample members are retired from the study after age 75. In 2006 the SDR had a response rate of 78%.
For further information on this report or the SDR, contact
Human Resources Statistics Program
Division of Science Resources Statistics
National Science Foundation
4201 Wilson Boulevard, Suite 965
Arlington VA 22230
 Thomas B. Hoffer is a principal research scientist, Karen Grigorian is a senior survey director, and Eric Hedberg is a research assistant at the National Opinion Research Center at the University of Chicago.
 Most recent doctorate recipients are defined as those who earned their degrees within the 5 years preceding the survey reference date.
 All differences stated in this InfoBrief are significant at the 95% level.
 The broad category life sciences includes doctorates earned in biological, agricultural, or environmental life sciences, as well as those earned in health fields.
 The time-since-doctorate variable was calculated by converting the number of months between receipt of the doctorate and the SDR reference date (1 April 2006) into decimal years.
 Among all doctorate recipients in 2006, 28% reported having had or currently working in just one postdoc, 8% reported having had two postdocs, and 2% three or more. Multiple postdocs were more often reported by life sciences (17%) and physical sciences (12%) doctorate recipients.
 National Science Foundation, Division of Science Resources Statistics. 1998. Has the Use of Postdocs Changed? Issue Brief NSF 99-310. Arlington, VA. Available at http://www.nsf.gov/statistics/issuebrf/ib99310.htm.
 Commission on Professionals in Science and Technology (CPST). 2002. Postdocs: What We Know and What We Would Like to Learn. Proceedings of an NSF/CPST Professional Societies Workshop, 4 December 2002. Washington, DC. Available at http://www.cpst.org/Postdoc.pdf. Accessed November 2007.
 Committee on Science, Engineering, and Public Policy, National Academies. 2000. Enhancing the Postdoctoral Experience for Scientists and Engineers. Washington, DC: National Academy Press. Available at http://www.nap.edu/books/0309069963/html/. Accessed November 2007.
 National Science Board (NSB). 1998. Science and Engineering Indicators: 1998. Vol. 1, chap. 3. NSB 98-01. Arlington, VA: National Science Foundation. Available at http://www.nsf.gov/statistics/seind98/.